Heating demand prediction method based on artificial intelligence

Abstract

The invention discloses a heating demand prediction method based on artificial intelligence, and relates to the technical field of building heating. The method comprises the steps that building characteristic environment data and heating heat data are collected, a heating heat significant correlation value and a correlation coefficient are obtained through chi-square testing and Pearson correlation analysis, a heating heat importance index is calculated, and the building characteristic environment data are screened. Inputting the screened data and the heating heat data into a support vector machine regression model for training; building characteristic environment data are analyzed based on the Fourier law and the Newton cooling law, a wall heat conduction loss value and a window heat convection loss value are calculated, and a building heat balance dynamic equation is established; building characteristic environment screening data are collected in real time and input into the support vector machine regression model, and a first heating heat prediction value is obtained; and synchronously inputting real-time building characteristic environment data into the dynamic equation to obtain a second heating heat prediction value. And fusing the two types of prediction values to generate a comprehensive prediction value to realize heating demand prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of building heating, and in particular to a heating demand prediction method based on artificial intelligence. Background Art

[0002] Heating demand forecasting plays an important role in building energy management, especially in energy conservation, emission reduction, and improving energy efficiency. With the impact of global climate change and fluctuating energy prices, accurate heating demand forecasting has become increasingly critical for controlling building energy consumption. In recent years, with the development of artificial intelligence and big data technologies, heating demand forecasting methods have gradually become more intelligent and sophisticated. Machine learning algorithms, particularly support vector machines and neural networks, have begun to be widely used in heating forecasting. These methods can handle complex nonlinear relationships and, by combining large amounts of real-time data such as meteorological data and building energy consumption data, improve the accuracy of forecasts. Furthermore, intelligent forecasting methods can dynamically adjust models based on real-time data to further optimize forecast results.

[0003] Existing machine learning methods often directly use original features for prediction in heating demand forecasting, but lack the ability to quantify the relative impact of each original feature on the heating heat supply. The prediction model contains redundant features, which leads to reduced accuracy in heating demand prediction. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the present invention provides a heating demand forecasting method based on artificial intelligence to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a heating demand prediction method based on artificial intelligence, comprising the following steps: Step S1: collecting building heating data and classifying it by time to obtain a building heating data set, wherein the building heating data set includes building characteristic environment data and heating heat data; Step S2: Calculate the heating heat correlation coefficient by performing a Pearson correlation analysis on the building characteristic environment data and the heating heat data; Calculate the heating heat significant correlation value by performing a chi-square test on the building characteristic environment data and the heating heat data; Step S3: Calculate the heating heat importance index by combining the heating heat correlation coefficient and the heating heat significant correlation value; perform feature screening on the building characteristic environment data using the heating heat importance index to obtain building characteristic environment screening data; Step S4: Inputting the building characteristic environment screening data and the heating heat data into a support vector machine regression model for training to obtain a trained support vector machine regression model; analyzing the building characteristic environment data using Fourier's law to obtain a building wall heat conduction loss value; analyzing the building characteristic environment data using Newton's law of cooling to obtain a building window heat convection loss value; combining the building wall heat conduction loss value and the building window heat convection loss value to establish a building thermal balance dynamic equation; Step S5: Collecting building characteristic environment screening data in real time and inputting it into the trained support vector machine regression model to obtain a first heating heat prediction value; at the same time, inputting the real-time collected building characteristic environment data into the building thermal balance dynamic equation to output a second heating heat prediction value; Step S6: combining the first heating heat prediction value and the second heating heat prediction value to calculate a comprehensive heating heat prediction value, thereby realizing the prediction of heating demand.

[0006] Preferably, the step of performing Pearson correlation analysis on the building characteristic environment data and the heating heat data to calculate the heating heat correlation coefficient comprises the following steps: By performing Pearson correlation analysis on building characteristic environment data and heating heat data, the heating heat correlation coefficient is calculated:

[0007] in, The first feature in the building characteristic environment data The heating heat correlation coefficient of each feature is The first feature in the building characteristic environment data The first feature data, The first feature in the building characteristic environment data The average value of the features, The first data, is the average value of heating heat data, Indicates the total number of samples.

[0008] Preferably, the method of calculating the significant correlation value of heating heat by performing a chi-square test on the building characteristic environment data and the heating heat data comprises the following specific steps: By statistically analyzing the building characteristic environmental data The number of samples under the heating heat label for each feature is used to create a contingency table:

[0009] in, Indicates the building characteristic environment data Contingency table of features and heating heat labels, Indicates the building characteristic environment data The first feature Category in The number of samples of heating heat labels, The first The total number of categories of features, The number of labels indicating the heating heat label; Calculate the building characteristic environment data DF The first feature Category in Expected frequency of heating heat labels:

[0010] in, The first character representing the characteristics in the building characteristic environment data The first feature Category in The expected frequency of heating heat labels, The first character representing the characteristics of building environment data The first feature Category in The number of samples of heating heat labels, The first character representing the characteristics of building environment data The first feature The index of the categories, Indicates the The index of the heating heat label, The first The total number of categories of features, The number of labels indicating the heating heat label; Calculate the chi-square statistic:

[0011] in, The first feature in the building characteristic environment data The chi-square statistic for each feature, The first character representing the characteristics of building environment data The first feature Category in The number of samples of heating heat labels, The first character representing the characteristics of building environment data The first feature The index of the categories, Indicates the The index of the heating heat label, The first The total number of categories of features, The number of labels indicating the heating heat label; Making a significance judgment: ,in, The first feature in the building characteristic environment data The degrees of freedom of the characteristics are found through the chi-square distribution critical value table. and the significance level is Chi-square distribution critical value ,when When the first feature in the building characteristic environment data The characteristics are significantly correlated with heating heat; By the characteristics of the building characteristic environment data The Cramer correlation coefficient of the characteristics is used to calculate the significant correlation value of heating heat:

[0012] in, The first feature in the building characteristic environment data The heating heat of each feature is significantly correlated with the value of The first feature in the building characteristic environment data The chi-square statistic for each feature, The degrees of freedom are and the significance level is Chi-square distribution critical value , The first character representing the characteristics of building environment data The first feature Category in The number of samples of heating heat labels, The first character representing the characteristics of building environment data The first feature The index of the categories, Indicates the The index of the heating heat label, The first The total number of categories of features, The number of tags representing the heating heat tags.

[0013] Preferably, the step of calculating the heating heat importance index by combining the heating heat correlation coefficient and the heating heat significant correlation value comprises the following steps: By combining the heating heat correlation coefficient and the heating heat significant correlation value, the heating heat importance index is calculated:

[0014] in, The first feature in the building characteristic environment data The heating heat importance index of each feature, The first feature in the building characteristic environment data The heating heat correlation coefficient of each feature is The first feature in the building characteristic environment data The heating heat of each feature is significantly correlated.

[0015] Preferably, the feature screening of the building characteristic environment data by the heating heat importance index to obtain the building characteristic environment screening data comprises the following specific steps: The building characteristic environment data is screened by the heating heat importance index. If the heating heat importance index of the feature is greater than or equal to the preset threshold, the first If the feature is less than the preset threshold, the first features, and finally obtain the building characteristic environment screening data.

[0016] Preferably, the analyzing of the building characteristic environment data by Fourier's law to obtain the building wall heat conduction loss value comprises the following specific steps: By analyzing the building characteristic environmental data using Fourier's law, the heat conduction loss value of the building wall is obtained:

[0017] in, is the wall heat conduction loss value, is the thermal conductivity of the wall material, which is determined by the material of the wall itself. is the heat transfer area of the wall, is the temperature inside the building, Temperature outside the building, is the wall thickness.

[0018] Preferably, the step of analyzing the building characteristic environment data using Newton's law of cooling to obtain the building window heat convection loss value comprises the following specific steps: By analyzing the building characteristic environmental data using Newton's law of cooling, the heat convection loss value of the building window is obtained:

[0019] in, Indicates the heat convection loss value of the building window, Indicates the window heat transfer coefficient, which is determined by the material of the window itself. represents the surface area of the window, is the temperature inside the building, Temperature outside the building.

[0020] Preferably, the step of establishing a dynamic equation of building thermal balance by combining the building wall heat conduction loss value and the building window heat convection loss value comprises the following specific steps: By combining the heat conduction loss value of the building wall and the heat convection loss value of the building window, a dynamic equation of building thermal balance is established:

[0021] in, for The building heating heat at the moment, is the building heat capacity, ,in is the density of building materials, is the building volume, is the specific heat capacity of building materials, The temperature inside the building is The rate of change at a moment, = , is the time interval, for The indoor temperature at the moment, for The indoor temperature at the moment, for The wall heat conduction loss value at the moment, express The heat convection loss value of the building window at the moment, for Solar radiation gain at time = ,in, is the solar radiation absorption rate of the window, which is determined by the material of the window itself. represents the surface area of the window, for The intensity of solar radiation at the time.

[0022] Preferably, the real-time collection of building characteristic environment screening data and inputting it into a trained support vector machine regression model to obtain a first heating heat prediction value; at the same time, the real-time collected building characteristic environment data is input into a building thermal balance dynamic equation to output a second heating heat prediction value, including the following specific steps: Collect building characteristic environment screening data in real time and input it into the trained support vector machine regression model to obtain the first heating heat prediction value:

[0023] in, for The first heating heat forecast value at the moment, represents the number of support vectors, and are all Lagrange multipliers of the support vector machine regression model, Indicates the The indices of the support vectors, is the kernel function, which calculates the input building characteristic environment screening data With the Support vectors , b is the bias term of the support vector machine regression model; At the same time, the real-time collected building characteristic environment data is input into the building heat balance dynamic equation, and the output is the second heating heat prediction value:

[0024] in, for The second heating heat forecast value at the moment, is the building heat capacity, ,in is the density of building materials, is the building volume, is the specific heat capacity of building materials, The temperature inside the building at time The rate of change under for The wall heat conduction loss value at the moment, is the solar radiation gain, express The heat convection loss value of the building window at the moment, = ,in, is the solar radiation absorption rate of the window, which is determined by the material of the window itself. represents the surface area of the window, for The intensity of solar radiation at the time.

[0025] Preferably, the method of calculating a comprehensive heating heat prediction value by combining the first heating heat prediction value and the second heating heat prediction value to achieve the prediction of the heating demand includes the following specific steps: By combining the first heating heat prediction value and the second heating heat prediction value, the comprehensive heating heat prediction value is calculated:

[0026] in, for The comprehensive forecast value of heating heat at the moment, for The first heating heat forecast value at the moment, for The second heating heat forecast value at the moment, is the sensitivity coefficient, which is used to adjust the speed of the impact of the support vector machine regression model accuracy. is the accuracy threshold, is the model accuracy of the support vector machine regression model, which is determined by the mean square error of the support vector machine regression model.

[0027] Beneficial effects: The present invention provides a heating demand forecasting method based on artificial intelligence, involving machine learning and deep learning technologies, which has the following beneficial effects: (1) By combining the heating heat correlation coefficient and the heating heat significant correlation value to calculate the heating heat importance index, it is possible to cover the association analysis of categorical variables and continuous variables with the help of multivariate statistical methods, avoiding the limitation of a single method on the data type. Among them, the significant correlation value quantifies the nonlinear correlation degree between the categorical feature and the heating demand, and the correlation coefficient measures the linear correlation strength of the continuous feature. The two together constitute a unified quantitative standard, enhance the comprehensiveness and reliability of feature evaluation, provide a multidimensional basis for feature screening, ensure the retention of key environmental parameters that have a significant impact on heating heat, and improve the prediction efficiency of subsequent models.

[0028] (2) The dynamic equation of building thermal balance is established by combining the heat conduction loss value of the building wall and the heat convection loss value of the building window. Based on physical principles such as Fourier's law and Newton's law of cooling, the conduction, convection effects and solar radiation gain in the building heat transfer process are incorporated into a unified framework to dynamically describe the relationship between the temperature change inside the building and the heating demand. This equation fully considers the properties of building materials such as thermal conductivity and specific heat capacity, structural parameters such as wall thickness and window area, and external environment such as indoor and outdoor temperature and solar radiation intensity, so that the prediction process has clear physical meaning and can provide mechanism-level heat demand analysis for different building types and climatic conditions, thereby enhancing the interpretability and scenario adaptability of the prediction results.

[0029] (3) Combine the first heating heat prediction value and the second heating heat prediction value to calculate the comprehensive prediction value, integrate the statistical learning ability of the data-driven model with the prior knowledge of the physical mechanism model, break through the limitations of the single modeling logic, and achieve high precision, strong robustness and scenario adaptability in heating demand prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0031] Figure 1 This is a flowchart of the steps of a heating demand prediction method based on artificial intelligence proposed by the present invention; Figure 2 This is a hierarchical diagram of the steps of the heating demand forecasting method based on artificial intelligence proposed in the present invention. DETAILED DESCRIPTION

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0033] See also Figure 1-Figure 2 , the present invention provides a technical solution: a heating demand prediction method based on artificial intelligence.

[0034] Step S1: collecting heating data of a building and classifying it by time to obtain a building heating data set, wherein the building heating data set includes building characteristic environment data and heating heat data.

[0035] The specific steps for collecting building heating data and classifying it by time to obtain a building heating dataset are as follows: First, multi-source data collection: including building characteristic environmental data and heating heat data. Building characteristic environmental data refers to meteorological data and building characteristic data. Meteorological data includes real-time external temperature, humidity, wind speed, precipitation, sunshine duration, etc., which cover climate characteristics of different seasons and daytimes. Building characteristic data covers physical properties such as building type (such as residential, office, and commercial), building volume, envelope parameters (wall material, thickness, thermal conductivity, window type, area, and insulation performance), material density, specific heat capacity, and insulation grade. Heating heat data refers to historical heating operation data, including heat supply in each time period. The collected data is cleaned and preprocessed: missing values are filled (such as through interpolation of adjacent time periods or statistical mean completion) and outlier detection (based on statistical methods or box plot identification and correction) to ensure data integrity and accuracy. Simultaneously, numerical data is normalized (such as Z-score normalization) to eliminate the impact of dimensional differences on subsequent analysis. Temporal Dimension Classification and Feature Engineering: For meteorological data and heating heat data, we classified features based on hourly, daily, and monthly time scales. Building feature data, due to its static nature, was directly incorporated into the dataset as a fixed feature. Finally, dataset integration was performed: the building characteristic environmental screening data and heating heat data were aligned and merged by timestamp to form a structured building heating dataset.

[0036] Step S2: By performing a Pearson correlation analysis on the building characteristic environment data and the heating heat data, the heating heat correlation coefficient is calculated; by performing a chi-square test on the building characteristic environment data and the heating heat data, the heating heat significant correlation value is calculated.

[0037] By performing Pearson correlation analysis on building characteristic environment data and heating heat data, the heating heat correlation coefficient is calculated:

[0038] in, The first feature in the building characteristic environment data The heating heat correlation coefficient of each feature is The first feature in the building characteristic environment data The first feature data, The first feature in the building characteristic environment data The average value of the features, The first data, is the average value of heating heat data, Indicates the total number of samples.

[0039] It should be noted that the Pearson correlation is a statistical method used to measure the strength and direction of the linear relationship between two continuous variables. In feature correlation analysis, the Pearson correlation coefficient can help determine whether there is a linear correlation between two features. The value of the Pearson correlation coefficient ranges from -1 to 1, where: 1 indicates a perfect positive correlation, and the two features increase or decrease accordingly; -1 indicates a perfect negative correlation, when one feature increases, the other feature decreases; 0 indicates no linear correlation. Pearson correlation is often used to analyze the relationship between numerical data. By calculating the Pearson correlation coefficient, it can help identify strong or weak correlations between features, thereby providing a basis for further data modeling and prediction.

[0040] Bin the heating heat data into discrete categories and obtain heating heat labels: high, medium, and low.

[0041] It should be noted that the heating heat data is divided into discrete categories to obtain heating heat labels. For example, if the heating heat , the heating heat label is low, , the heating heat label is medium, , the heating heat label is high.

[0042] Chi-square tests were performed on the building characteristic environment data and the heating heat data after discretization.

[0043] It should be noted that the chi-square test is a statistical test method primarily used to test relationships or differences between categorical data. In feature association analysis, the chi-square test is primarily used to test whether there is a significant association between two categorical variables. By comparing the observed frequencies with the expected frequencies, the chi-square test can help determine whether two variables are independent or interdependent. If there is no association between the two variables, the chi-square value is typically small, close to zero; if there is a significant association between them, the chi-square value is large, indicating that the observed data differs significantly from the hypothesized expected value. Simply put, the chi-square test can reveal whether there is a statistical correlation between features and is widely used in fields such as market analysis, social surveys, and medical research to help analyze the relationships between different factors.

[0044] By statistically analyzing the building characteristic environmental data The number of samples under the heating heat label for each feature is used to create a contingency table:

[0045] in, Indicates the building characteristic environment data Contingency table of features and heating heat labels, Indicates the building characteristic environment data The first feature Category in The number of samples of heating heat labels, The first The total number of categories of features, The number of tags representing the heating heat tags.

[0046] It should be noted that the building characteristic environment data The first feature Category, building characteristic environment data If the feature is a primitive categorical variable, such as "building type" = residential / commercial / office, "insulation level" = Class A / Class B / Class C, it can be used directly; If a feature is a continuous variable (such as external temperature, wall thickness), it must first be converted into a categorical variable through discretization methods, such as dividing it into fixed intervals (such as "external temperature" = low temperature / medium temperature / high temperature).

[0047] Calculate the building characteristic environment data DF The first feature Category in Expected frequency of heating heat labels:

[0048] in, The first character representing the characteristics in the building characteristic environment data The first feature Category in The expected frequency of heating heat labels, The first character representing the characteristics of building environment data The first feature Category in The number of samples of heating heat labels, The first character representing the characteristics of building environment data The first feature The index of the categories, Indicates the The index of the heating heat label, The first The total number of categories of features, The number of tags representing the heating heat tags.

[0049] Calculate the chi-square statistic:

[0050] in, The first feature in the building characteristic environment data The chi-square statistic for each feature, The first character representing the characteristics of building environment data The first feature Category in The number of samples of heating heat labels, The first character representing the characteristics of building environment data The first feature The index of the categories, Indicates the The index of the heating heat label, The first The total number of categories of features, The number of tags representing the heating heat tags.

[0051] Making a significance judgment: ,in, The first feature in the building characteristic environment data The degrees of freedom of the characteristics are found through the chi-square distribution critical value table. and the significance level is Chi-square distribution critical value ,when When the first feature in the building characteristic environment data The characteristics are significantly correlated with heating heat. , then the first feature in the building characteristic environment data There was no significant correlation between these characteristics and heating quantity.

[0052] When necessary, the significance level α is a core concept in hypothesis testing in statistics, used to quantify the maximum risk of a researcher accepting a "false rejection of the null hypothesis." It represents the probability of mistakenly concluding that a significant association exists when the null hypothesis (e.g., "variables are unrelated") is actually true. For example, if α is set to 0.05, this means that an average of up to five "false positive" conclusions (i.e., falsely concluding that there is an association when there is none) are allowed in 100 independent tests. The choice of this threshold is usually disciplinary convention (e.g., social sciences often use α = 0.05, while high-precision fields such as particle physics may require α = 0.0000003). Its essence is to control the risk of Type I error (the error of rejecting the true positive).

[0053] By the characteristics of the building characteristic environment data The Cramer correlation coefficient of the characteristics is used to calculate the significant correlation value of heating heat:

[0054] in, The first feature in the building characteristic environment data The heating heat of each feature is significantly correlated with the value of The first feature in the building characteristic environment data The chi-square statistic for each feature, The degrees of freedom are and the significance level is Chi-square distribution critical value , The first character representing the characteristics of building environment data The first feature Category in The number of samples of heating heat labels, The first character representing the characteristics of building environment data The first feature The index of the categories, Indicates the The index of the heating heat label, The first The total number of categories of features, The number of tags representing the heating heat tags.

[0055] Step S3: combining the heating heat correlation coefficient and the heating heat significant association value to calculate the heating heat importance index; performing feature screening on the building characteristic environment data using the heating heat importance index to obtain building characteristic environment screening data.

[0056] By combining the heating heat correlation coefficient and the heating heat significant correlation value, the heating heat importance index is calculated:

[0057] in, The first feature in the building characteristic environment data The heating heat importance index of each feature, The first feature in the building characteristic environment data The heating heat correlation coefficient of each feature is The first feature in the building characteristic environment data The heating heat of each feature is significantly correlated.

[0058] It should be noted that when conducting the chi-square test, if the first The characteristics have no significant correlation with heating heat, namely , the default Heating heat importance index of each feature Equal to 0.

[0059] It should be noted that the heating heat importance index is calculated by combining the heating heat correlation coefficient and the heating heat significant correlation value, and the correlation analysis of categorical variables and continuous variables is covered with the help of multivariate statistical methods, avoiding the limitations of a single method on data types, and enhancing the comprehensiveness and reliability of feature evaluation through double verification. At the same time, a unified quantitative standard is constructed to support feature importance ranking; among them, the heating heat significant correlation value is a quantitative reflection of the correlation strength of categorical features in the importance index, which directly reflects the nonlinear correlation between categorical variables (such as building type, insulation level) and heating demand, and provides the contribution of the categorical dimension to the importance index; and the heating heat correlation coefficient is the core indicator of the linear correlation strength of continuous features in the importance index. By quantifying the linear correlation between continuous variables (such as external temperature, wall thermal conductivity) and heating heat, it lays the foundation for the evaluation of the continuous dimension of the importance index. The two together constitute the calculation core of the importance index, realizing the multi-dimensional value mining of building characteristic environmental data.

[0060] The building characteristic environment data is screened by the heating heat importance index, and the first If the heating heat importance index of the feature is greater than or equal to the preset threshold, the first If the feature is less than the preset threshold, the first features, and finally obtain the building characteristic environment screening data.

[0061] Step S4: Input the building characteristic environment screening data and heating heat data into a support vector machine regression model for training to obtain a trained support vector machine regression model; analyze the building characteristic environment data using Fourier's law to obtain the building wall heat conduction loss value; analyze the building characteristic environment data using Newton's cooling law to obtain the building window heat convection loss value; combine the building wall heat conduction loss value and the building window heat convection loss value to establish a building thermal balance dynamic equation.

[0062] After data standardization, the building characteristic environmental screening data and heating heat data are used as independent variables and the heating heat data as dependent variables. Both the building characteristic environmental screening data and the heating heat data are divided into training and test sets in a ratio of 7:3 and input into the support vector machine regression model for training. The radial basis kernel function (RBF) is preferentially selected for the support vector machine regression model. The regularization parameter C and the kernel function parameter γ of the support vector machine regression model are combined and optimized using the grid search method. Cross-validation is also used to evaluate the model's generalization ability to avoid overfitting. During the training process, the trend of the model's mean squared error (MSE) is monitored. Training is terminated when the error no longer decreases significantly after multiple rounds of iterations. Finally, the model's prediction performance is quantified using indicators such as the mean absolute error (MAE) and root mean square error (RMSE) of the test set to ensure that the trained support vector machine regression model has reliable generalization ability.

[0063] By analyzing the building characteristic environmental data using Fourier's law, the heat conduction loss value of the building wall is obtained:

[0064] in, is the wall heat conduction loss value, is the thermal conductivity of the wall material, which is determined by the material of the wall itself. is the heat transfer area of the wall, is the temperature inside the building, Temperature outside the building, is the wall thickness.

[0065] It should be noted that Fourier's law is a fundamental physical law in the field of heat conduction, used to quantify the rate of heat transfer through solid materials. Its core reveals that under steady-state conditions, the heat flux through a homogeneous material per unit time is proportional to the temperature gradient on both sides of the material and the cross-sectional area perpendicular to the direction of heat flow, and inversely proportional to the thickness of the material in the direction of heat flow. The proportionality constant is the thermal conductivity of the material, which is an inherent physical property of the material that characterizes its ability to conduct heat. The higher the thermal conductivity, such as metal, the stronger the thermal conductivity of the material; the lower the thermal conductivity, such as insulation foam, the better the thermal insulation of the material.

[0066] By analyzing the building characteristic environmental data using Newton's law of cooling, the heat convection loss value of the building window is obtained:

[0067] in, Indicates the heat convection loss value of the building window, Indicates the window heat transfer coefficient, which is determined by the material of the window itself. represents the surface area of the window, is the temperature inside the building, Temperature outside the building.

[0068] It's important to note that Newton's law of cooling is a fundamental physical law that describes the phenomenon of convective heat transfer between a fluid and a solid surface. It is used to quantify the rate of heat transfer due to air flow. Its core principle states that the amount of heat lost by surface convection per unit time is proportional to the temperature difference between the solid surface and the surrounding fluid, as well as the surface area involved in the heat transfer. The proportionality constant is called the convective heat transfer coefficient, which comprehensively reflects the influence of fluid properties (such as air density, viscosity, and specific heat capacity), flow state (natural or forced convection), and surface characteristics (such as roughness and orientation) on the intensity of heat transfer. A higher heat transfer coefficient (such as a window in strong wind) results in faster convective heat dissipation; conversely, a lower heat transfer coefficient (such as a window in still air) results in slower heat dissipation.

[0069] By combining the heat conduction loss value of the building wall and the heat convection loss value of the building window, a dynamic equation of building thermal balance is established:

[0070] in, for The building heating heat at the moment, is the building heat capacity, ,in is the density of building materials, is the building volume, is the specific heat capacity of building materials, The temperature inside the building is The rate of change at a moment, = , is the time interval, for The indoor temperature at the moment, for -1 hour indoor temperature, for The wall heat conduction loss value at the moment, express The heat convection loss value of the building window at the moment, for Solar radiation gain at time = ,in, is the solar radiation absorption rate of the window, which is determined by the material of the window itself. represents the surface area of the window, for The intensity of solar radiation at the time.

[0071] Step S5: Collect building characteristic environment screening data in real time and input it into the trained support vector machine regression model to obtain a first heating heat prediction value; at the same time, input the real-time collected building characteristic environment data into the building thermal balance dynamic equation and output a second heating heat prediction value.

[0072] Collect building characteristic environment screening data in real time and input it into the trained support vector machine regression model to obtain the first heating heat prediction value:

[0073] in, for The first heating heat forecast value at the moment, represents the number of support vectors, and are all Lagrange multipliers of the support vector machine regression model, Indicates the The indices of the support vectors, is the kernel function, which calculates the input building characteristic environment screening data With the Support vectors , and b is the bias term of the support vector machine regression model.

[0074] At the same time, the real-time collected building characteristic environment data is input into the building heat balance dynamic equation, and the output is the second heating heat prediction value:

[0075] in, for The second heating heat forecast value at the moment, is the building heat capacity, ,in is the density of building materials, is the building volume, is the specific heat capacity of building materials, The temperature inside the building is The rate of change at a moment, for The wall heat conduction loss value at the moment, is the solar radiation gain, express The heat convection loss value of the building window at the moment, = ,in, is the solar radiation absorption rate of the window, which is determined by the material of the window itself. represents the surface area of the window, for The intensity of solar radiation at the time.

[0076] It should be noted that the real-time collected building characteristic environment data is input into the building heat balance dynamic equation, and the output is the second heating heat prediction value. For example: the heat capacity C is 1000kJ / ℃, and the indoor temperature needs to rise from 19℃ to 20℃, then =1℃ / h, then ==1000kw, if The total during the change process is 50kw, so the second heating heat prediction value is 1050kw.

[0077] Step S6: By combining the first heating heat prediction value and the second heating heat prediction value, a comprehensive heating heat prediction value is calculated to achieve the prediction of heating demand.

[0078] By combining the first heating heat prediction value and the second heating heat prediction value, the comprehensive heating heat prediction value is calculated:

[0079] in, for The comprehensive forecast value of heating heat at the moment, for The first heating heat forecast value at the moment, for The second heating heat forecast value at the moment, is the sensitivity coefficient, which is used to adjust the speed of the impact of the support vector machine regression model accuracy. is the accuracy threshold, is the model accuracy of the support vector machine regression model, which is determined by the mean square error of the support vector machine regression model.

[0080] It should be noted that the comprehensive prediction value is calculated by combining the first heating heat prediction value (output of the support vector machine regression model) and the second heating heat prediction value (output of the building heat balance dynamic equation). is the weight function of the first heating heat prediction value, which is a Sigmoid function, and the mean square error of the support vector machine regression model in the test set is As input, to achieve dynamic weight allocation: When the support vector machine regression model has high accuracy, When it approaches 1, it means that the data-driven support vector machine regression model is trusted first. When the accuracy of the support vector machine regression model is low, If the prediction approaches zero, the physical model of the building's thermal balance dynamic equation serves as a safeguard. By integrating the statistical learning capabilities of data-driven models with the prior knowledge of physical mechanism models, the limitations of single modeling logic can be overcome, achieving high-precision, robust, and scenario-adaptive heating demand forecasts. This overcomes the inherent flaws of single models. Limitations of data-driven models: While machine learning models such as support vector machines can mine nonlinear correlations through historical data, they rely on data coverage and quality. In extreme climates (such as cold snaps and abnormally high temperatures) or new building types (such as passive energy-saving buildings), a lack of sufficient historical samples can easily lead to data blind spots and prediction bias. Limitations of physical mechanism models: While the thermal balance dynamic equation is based on physical principles such as Fourier's law and Newton's law of cooling and can capture the essential laws of heat conduction and convection, it lacks the ability to capture real-time environmental variables (such as indoor temperature fluctuations caused by user behavior and non-steady-state meteorological conditions) and dynamic changes in building parameters (such as thermal conductivity drift due to wall aging). By combining the two types of models, machine learning is used to fill the gaps in the physical model's description of complex nonlinear relationships. At the same time, the physical model provides mechanism constraints for machine learning to avoid it falling into the data fitting trap, especially to improve prediction reliability when data is sparse or the scene changes suddenly.

[0081] This paper proposes an artificial intelligence-based heating demand forecasting method that achieves accurate predictions through a multi-step fusion of statistical data analysis and physical models. First, building heating data is collected and categorized by time. A dataset is constructed through data cleaning and feature engineering. Then, the significant association value and correlation coefficient of heating heat are calculated through chi-square tests and Pearson correlation analysis. The importance index is combined to filter features, and this filtered data is fed into a support vector machine regression model for training. A dynamic building heat balance equation is established based on Fourier's law and Newton's law of cooling. Finally, the prediction results of the machine learning model and the physical model are integrated to achieve a comprehensive forecast of heating demand.

[0082] By combining the heating heat correlation coefficient and the heating heat significant correlation value to calculate the heating heat importance index, multivariate statistical methods can be used to cover the association analysis of categorical and continuous variables, avoiding the limitations of a single method on the data type. The significant correlation value quantifies the degree of nonlinear association between categorical features and heating demand, while the correlation coefficient measures the strength of the linear association of continuous features. Together, they form a unified quantitative standard, enhancing the comprehensiveness and reliability of feature evaluation, providing a multidimensional basis for feature screening, ensuring the retention of key environmental parameters that significantly affect heating heat, and improving the predictive performance of subsequent models.

[0083] A dynamic building heat balance equation is established by combining heat conduction losses through building walls and heat convection losses through building windows. Based on physical principles such as Fourier's law and Newton's law of cooling, this equation incorporates conduction, convection, and solar radiation gain in the building's heat transfer process into a unified framework, dynamically describing the relationship between internal building temperature changes and heating demand. This equation fully considers building material properties (such as thermal conductivity and specific heat capacity), structural parameters (such as wall thickness and window area), and the external environment (such as indoor and outdoor temperatures and solar radiation intensity), giving the prediction process clear physical meaning. It can provide mechanistic heat demand analysis for different building types and climate conditions, enhancing the interpretability and adaptability of the prediction results.

[0084] The comprehensive prediction value is calculated by combining the first heating heat prediction value (output of the support vector machine regression model) and the second heating heat prediction value (output of the building heat balance dynamic equation). This integrates the statistical learning ability of the data-driven model with the prior knowledge of the physical mechanism model, breaking through the limitations of a single modeling logic and achieving high precision, strong robustness and scenario adaptability in heating demand prediction.

[0085] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include," "comprise," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations. The phrase "includes an element defined by..." does not exclude the presence of other identical elements in the process, method, article, or device that includes the element.

[0086] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A heating demand forecasting method based on artificial intelligence, characterized by: The following steps are involved: Step S1: collecting building heating data and classifying it by time to obtain a building heating data set, wherein the building heating data set includes building characteristic environment data and heating heat data; Step S2: Calculate the heating heat correlation coefficient by performing a Pearson correlation analysis on the building characteristic environment data and the heating heat data; Calculate the heating heat significant correlation value by performing a chi-square test on the building characteristic environment data and the heating heat data; Step S3: Calculate the heating heat importance index by combining the heating heat correlation coefficient and the heating heat significant correlation value; perform feature screening on the building characteristic environment data using the heating heat importance index to obtain building characteristic environment screening data; Step S4: Inputting the building characteristic environment screening data and the heating heat data into a support vector machine regression model for training to obtain a trained support vector machine regression model; analyzing the building characteristic environment data using Fourier's law to obtain a building wall heat conduction loss value; analyzing the building characteristic environment data using Newton's law of cooling to obtain a building window heat convection loss value; combining the building wall heat conduction loss value and the building window heat convection loss value to establish a building thermal balance dynamic equation; Step S5: Collecting building characteristic environment screening data in real time and inputting it into the trained support vector machine regression model to obtain a first heating heat prediction value; at the same time, inputting the real-time collected building characteristic environment data into the building thermal balance dynamic equation to output a second heating heat prediction value; Step S6: combining the first heating heat prediction value and the second heating heat prediction value to calculate a comprehensive heating heat prediction value, thereby realizing the prediction of heating demand.

2. The method for predicting heating demand based on artificial intelligence according to claim 1, characterized in that: The method of performing Pearson correlation analysis on the building characteristic environment data and the heating heat data to calculate the heating heat correlation coefficient includes the following steps: By performing Pearson correlation analysis on building characteristic environment data and heating heat data, the heating heat correlation coefficient is calculated: ； in, The first feature in the building characteristic environment data The heating heat correlation coefficient of each feature is The first feature in the building characteristic environment data The first feature data, The first feature in the building characteristic environment data The average value of the features, The first data, is the average value of heating heat data, Indicates the total number of samples.

3. The method for predicting heating demand based on artificial intelligence according to claim 2, characterized in that: The method of calculating the significant correlation value of heating heat by performing a chi-square test on the building characteristic environment data and the heating heat data includes the following specific steps: By statistically analyzing the building characteristic environmental data The number of samples under the heating heat label for each feature is used to create a contingency table: ； in, Indicates the building characteristic environment data Contingency table of features and heating heat labels, Indicates the building characteristic environment data The first feature Category in The number of samples of heating heat labels, The first The total number of categories of features, The number of labels indicating the heating heat label; Calculate the building characteristic environment data DF The first feature Category in Expected frequency of heating heat labels: ； in, The first character representing the characteristics in the building characteristic environment data The first feature Category in The expected frequency of heating heat labels, The first character representing the characteristics of building characteristic environment data The first feature Category in The number of samples of heating heat labels, The first character representing the characteristics of building environment data The first feature The index of the categories, Indicates the The index of the heating heat label, The first The total number of categories of features, The number of labels indicating the heating heat label; Calculate the chi-square statistic: ； in, The first feature in the building characteristic environment data The chi-square statistic for each feature, The first character representing the characteristics of building environment data The first feature Category in The number of samples of heating heat labels, The first character representing the characteristics of building environment data The first feature The index of the categories, Indicates the The index of the heating heat label, The first The total number of categories of features, The number of labels indicating the heating heat label; Making a significance judgment: ,in, The first feature in the building characteristic environment data The degrees of freedom of the characteristics are found through the chi-square distribution critical value table. and the significance level is Chi-square distribution critical value ,when When the first feature in the building characteristic environment data The characteristics are significantly correlated with heating heat; By the characteristics of the building characteristic environment data The Cramer correlation coefficient of the characteristics is used to calculate the significant correlation value of heating heat: ； in, The first feature in the building characteristic environment data The heating heat of each feature is significantly correlated with the value of The first feature in the building characteristic environment data The chi-square statistic for each feature, The degrees of freedom are and the significance level is Chi-square distribution critical value , The first character representing the characteristics of building environment data The first feature Category in The number of samples of heating heat labels, The first character representing the characteristics of building environment data The first feature The index of the categories, Indicates the The index of the heating heat label, The first The total number of categories of features, The number of tags representing the heating heat tags.

4. The method for predicting heating demand based on artificial intelligence according to claim 3, characterized in that: The calculation of the heating heat importance index by combining the heating heat correlation coefficient and the heating heat significant correlation value comprises the following steps: By combining the heating heat correlation coefficient and the heating heat significant correlation value, the heating heat importance index is calculated: ； in, The first feature in the building characteristic environment data The heating heat importance index of each feature, The first feature in the building characteristic environment data The heating heat correlation coefficient of each feature is The first feature in the building characteristic environment data The heating heat of each feature is significantly correlated.

5. The method for predicting heating demand based on artificial intelligence according to claim 4, characterized in that: The method of performing feature screening on the building characteristic environment data by using the heating heat importance index to obtain the building characteristic environment screening data comprises the following specific steps: The building characteristic environment data is screened by the heating heat importance index, and the first If the heating heat importance index of the feature is greater than or equal to the preset threshold, the first If the feature is less than the preset threshold, the first features, and finally obtain the building characteristic environment screening data.

6. The method for predicting heating demand based on artificial intelligence according to claim 5, characterized in that: The method of analyzing the building characteristic environment data by Fourier's law to obtain the building wall heat conduction loss value includes the following specific steps: By analyzing the building characteristic environmental data using Fourier's law, the heat conduction loss value of the building wall is obtained: ； in, is the wall heat conduction loss value, is the thermal conductivity of the wall material, which is determined by the material of the wall itself. is the heat transfer area of the wall, is the temperature inside the building, Temperature outside the building, is the wall thickness.

7. The method for predicting heating demand based on artificial intelligence according to claim 6, characterized in that: The method of analyzing the building characteristic environment data by using Newton's law of cooling to obtain the building window heat convection loss value includes the following specific steps: By analyzing the building characteristic environmental data using Newton's law of cooling, the heat convection loss value of the building window is obtained: ； in, Indicates the heat convection loss value of the building window, Indicates the window heat transfer coefficient, which is determined by the material of the window itself. represents the surface area of the window, is the temperature inside the building, Temperature outside the building.

8. The method for predicting heating demand based on artificial intelligence according to claim 7, characterized in that: The method of establishing a building thermal balance dynamic equation by combining the building wall heat conduction loss value and the building window heat convection loss value includes the following specific steps: By combining the heat conduction loss value of the building wall and the heat convection loss value of the building window, a dynamic equation of building thermal balance is established: ； in, for The building heating heat at the moment, is the building heat capacity, ,in is the density of building materials, is the building volume, is the specific heat capacity of building materials, The temperature inside the building is The rate of change at a moment, = , is the time interval, for The indoor temperature at the moment, for The indoor temperature at the moment, for The wall heat conduction loss value at the moment, express The heat convection loss value of the building window at the moment, for Solar radiation gain at time = ,in, is the solar radiation absorption rate of the window, which is determined by the material of the window itself. represents the surface area of the window, for The intensity of solar radiation at the time.

9. The method for predicting heating demand based on artificial intelligence according to claim 8, characterized in that: The real-time collection of building characteristic environment screening data and inputting it into a trained support vector machine regression model to obtain a first heating heat prediction value; at the same time, the real-time collected building characteristic environment data is input into a building thermal balance dynamic equation to output a second heating heat prediction value, including the following specific steps: Collect building characteristic environment screening data in real time and input it into the trained support vector machine regression model to obtain the first heating heat prediction value: ； in, for The first heating heat forecast value at the moment, represents the number of support vectors, and are all Lagrange multipliers of the support vector machine regression model, Indicates the The indices of the support vectors, is the kernel function, which calculates the input building characteristic environment screening data With the Support vectors , b is the bias term of the support vector machine regression model; At the same time, the real-time collected building characteristic environment data is input into the building heat balance dynamic equation, and the output is the second heating heat prediction value: ； in, for The second heating heat forecast value at the moment, is the building heat capacity, ,in is the density of building materials, is the building volume, is the specific heat capacity of building materials, The temperature inside the building at time The rate of change under for The wall heat conduction loss value at the moment, is the solar radiation gain, express The heat convection loss value of the building window at the moment, = ,in, is the solar radiation absorption rate of the window, which is determined by the material of the window itself. represents the surface area of the window, for The intensity of solar radiation at the time.

10. The method for predicting heating demand based on artificial intelligence according to claim 9, characterized in that: The method of calculating a comprehensive heating heat prediction value by combining the first heating heat prediction value and the second heating heat prediction value to achieve the prediction of heating demand includes the following specific steps: By combining the first heating heat prediction value and the second heating heat prediction value, the comprehensive heating heat prediction value is calculated: ； in, for The comprehensive forecast value of heating heat at the moment, for The first heating heat forecast value at the moment, for The second heating heat forecast value at the moment, is the sensitivity coefficient, which is used to adjust the speed of the impact of the support vector machine regression model accuracy. is the accuracy threshold, is the model accuracy of the support vector machine regression model, which is determined by the mean square error of the support vector machine regression model.

Images